ROTATIONALLY DISORDERED ILLITE/SMECTITE IN PALEOZOIC …clays.org/journal/archive/volume 43/43-3-271.pdf · Vol. 43, No. 3, 1995 RotationaUy disordered illite/smectite in Paleozoic

Clays and Clay Minerals, Vol. 43, No. 3, 271-284, 1995.

ROTATIONALLY DISORDERED ILLITE/SMECTITE IN PALEOZOIC K-BENTONITES

DOUGLAS K. McCARTY AND R. C. REYNOU3S JR.

Department of Earth Sciences, Dartmouth College, Hanover, New Hampshire 03755

Abstract- The three-dimensional crystal structure of rotationally disordered illite/smectite (I/S) in K-bentonite samples from the Appalachian basin and neighboring areas is described using the parameters of 1) Po, the proportion of zero-degree layer stacking rotations, such as in the polytype series IMd- IM; 2) Pcv, the proportion of 2:1 layers with cis-vacant (cv) octahedral sites that are randomly interstratified with trans-vacant (tv) layers; and 3) Pro, the proportion of layers with n- 60 ~ rotations (as opposed to n. 120 ~ in the rotated layers. These parameters were determined by computer modeling of experimental randomly oriented powder X-ray diffraction patterns.

The proportion of cv interstratification in the I/S increases with A1 and decreases with Mg and Fe content. The proportion of n. 60 ~ rotations in the rotated layers increases with Mg and Fe content. The cv 120 ~ disordered structure correlates with tetrahedral A 1 for Si substitution and increasing tetrahedral charge. The tv n.60 ~ disordered structures correlate with octahedral Mg for A1 substitution. The data indicate that the type of unit cell and nature of rotational disorder in I/S is controlled by the octahedral Mg content. The three-dimensional structures do not show any systematic correlation with Reichweite and percent expandability as determined from diffraction patterns of oriented sample preparations. Key Words--Bentonite, Cis-vacant octahedra, Crystal structure, Illite/Smectite, Polytype, Rotational disorder, X-ray diffraction.

I N T R O D U C T I O N

A number of papers describe the calculation of three- dimensional powder X-ray diffraction patterns ofphyl- losilicates with various kinds and amounts of disorder (e.g., Plan~on and Tchoubar 1977a, 1977b, Plan~on 1981, Plan~on et al 1988, Drits et a l 1984, Sakharov et a l 1990). The book by Drits and Tchoubar (1990) is a comprehensive state-of-the-art treatment of the subject. Reynolds (1993) developed a computer algo- ri thm (WILDFIRE) similar to those described above for calculating three-dimensional powder diffraction patterns of iUite and mixed-layer illite smectite (I/S) with various types and quantities of rotational disorder. This computer program was used to model diffraction patterns of I/S separated from 56 K-bentonite samples in and around the Appalachian basin (Figure I), thus quantifying the type and amount of crystallographic disorder in each sample. Such disordered I/S structures would have previously been called simply 1 M d polytypes using routine analytical methods. The goals of this study were to accurately model and char- acterize the three-dimensional I/S structures and to correlate compositional trends with the type of layer rotations and the distribution of octahedral cations.

An important distinction is made between 1) rotational disorder which refers to the angular displace- ments between the -X direction of two adjacent unit cells stacked along c and 2) Reichweite which describes the random or ordered stacking sequences of different kinds of phyllosilicate layers along Z in a MacEwan

Copyright �9 1995, The Clay Minerals Society

crystallite. In this paper the term disorder refers to rotational disorder unless otherwise specified.

Early work on illite polytypes by Maxwell and Hower (1967) recognized a general transformation from 1 M d to a 2 M structure which was complete when temperatures reached the biotite isograd. Experimental work with muscovite compositions indicated the general transformation 1 M d --+ 1 M --, 2 M I with increasing temperature (Yoder and Eugster 1955, Velde 1965). Work by Radoslovich and Norrish (1962) and Giiven (197 l) indicated that compositional variations affect what mica polytype results under given physical conditions.

The 2:1 layers in the usual dioctahedral mica structure have vacant M 1 (trans) octahedra, whereas both the M2 (cis) octahedral sites are occupied (Figure 2). This t rans-s i te vacancy and cation arrangement results in the mirror plane symmetry characteristic of the C 2 / m space group. The existence of 2: l phyllosilicates with different distributions of octahedral cations was first reported by Mrring and Oberlin (1967) for a sample of Wyoming montmoril lonite. Oblique texture electron diffraction (OTED) work by Tsipursky and Drits (1984) showed that dioctahedral smectites have a wide range ofoctahedral cation distributions over t rans and cis sites. The t r a m - o c c u p i e d (cis-vacant) unit cell is noncentrosymmetric and belongs to space group C2 (Tsipursky and Drits 1984) rather than C 2 / m , which applies to the t r a m - v a c a n t structure.

Specific distortions of cis and t rans octahedra result from the distribution of the octahedral cations and the

271

272 McCarty and Reynolds Clays and Clay Minerals

Figure 1. Sample locations and numbers.

resulting electrostatic forces. In the centrosymmetric 2:1 layer the larger vacant t rans octahedral site causes the center of the ditrigonal ring of the top tetrahedral sheet to be shifted from the center of the bottom tetrahedral sheet by more than the ideal -a /3 value. This type of distortion for A 1, Mg micaceous 2:1 phyllosili- cares leads to a monoclinic angle (fl) of about 101.3 ~ (Bailey 1984, Drits et a11993 , Reynolds and Thomson 1993). In the noncentrosymmetric 2:1 layer there is an undershift of the top tetrahedral sheet resulting in fl = 99.13 ~ (Tsipursky and Drits 1984, Drits et a l 1984, 1993). Diffraction effects are different for cis- and trans- vacant structures, for interstratification of layers with vacant cis and t rans sites, and also for structures with a statistical distribution of cations between cis and trans sites within single layers (Tsipursky and Drits 1984, Drits et a! 1984, 1993).

Experimental work by Reynolds (1992) showed that d ( h k l ) is independent of the average crystal thickness in the c direction for a mixed-layered I/S and there is no three-dimensional coherence across the expandable interfaces that separate illite layers or stacks of thin illite packets. A randomly oriented preparation of I/S diffracts in three dimensions like a randomly oriented powder of thin illite crystals. Reynolds (1993) demonstrated that c i s -vacan t (cv) structures exist in I/S exhibiting the 1 M and 1 M d stacking sequences, and that cv 2:1 layers can be interstratified with t rans -vacant (tv) 2:1 layers. Reynolds (1993) also demonstrated that the computer program WILDFIRE can success- fully model random powder diffraction patterns and

[]

Mirror Mane M(I)

K M(2) M(2)

I

F b !

Tv

Cv

Figure 2. Diagrammatic projection of (a) trans-vacant I/S structure; (b) cis-vacant I/S structure. Arrows show direction oftetrahedral rotation, solid circles are filled octahedral sites, boxes are vacant sites.

quantify the percent o f c i s - and t rans -vacan t octahedral sites along with the percent and degree of rotational disorder. Such crystallographic characterization has been accomplished for I/S from the Ordovician and Devonian bentonite beds in and around the Appala- chian basin and is the subject of this study.

Clay separations from Appalachian basin Ordovi- cian and Devonian K-bentonites were chosen for this study because I/S from K-bentonites are better suited to show the diffraction effects produced by different types and percentages of disorder and octahedral cation distributions than are I/S from shales and sandstones which almost always have diffraction peak interferences from other fine-grained minerals.

GEOLOGIC SETTING

K-ben ton i t e s

Ordovician volcanic ash layers deposited in epeiric seas presently cover about 1.3 million km 2 of eastern North America, distributed throughout thick sequences of carbonate and clastic rocks (Kolata et al 1984). Most of the Ordovician K-bentonite samples in this study were collected from the Deicke and MiUbrig and stratigraphic equivalent beds (Kolata et al 1984), the

Vol. 43, No. 3, 1 9 9 5 RotationaUy disordered illite/smectite in Paleozoic K-bentonites 273

Black River Group (Johnsson 1984), and the Utica Format ion (Mitchell 1992).

The Tioga Bentonite is the best known and most wide-spread volcanic ash layer in the middle Devonian rocks o f the Appalachian basin and probably covered all o f the northeastern Uni ted States during middle Devonian t ime (Dennison and Textoris 1970, Droste and Vitil iano 1973). The Devonian K-bentoni te samples in this study are from the Tioga or Tioga equivalent bentonite beds.

Post-deposition

A number o f workers have suggested a "fluid migration hypothesis" to explain post-deposit ional features in the Appalachian basin (e.g., Oliver 1986, Hearn et al 1987, Bethke and Marshak 1990). For example, I/S diagenesis in the distal Appalachian basin cannot be explained by the " t radi t ional" model of burial diagenesis (e.g., Perry and Hower 1970, Hower et al 1976, Jennings and Thompson 1986, McCarty and Thomp- son 1991) because conodont alteration index (CAI) values are too low (Harris 1979), and apparent burial depths too shallow. Yet, the I/S clays in Appalachian basin K-bentonites are R>__I structures with low to medium expandabil i ty similar to those found in deeply buried diagenetic sequences, and most o f these clays occur in geologic areas which show no evidence o f being subjected to the burial condit ions and associated temperatures that are considered necessary to produce this type of US (Huff and Tiirkmenoglu 1981, Elliott and Aronson 1987).

The I/S in the K-bentoni te beds from the Cincinnati Arch area contain about 70 to 80% illite layers (Huff and Tiirkmenoglu 1981, Elliott and Aronson 1993, results from this study). The max imum burial temperatures near 80~ (Harris 1979) appear to be less than the min imum temperature of about 100~ pro- posed by Hower et al (1976) which is needed to form ordered I/S of this type. Many of the K-bentoni tes are enclosed in carbonate strata which could not have pro- vided potassium, so an external source is indicated but not known. Roden et al (1992, 1993) found partially annealed zircon fission tracks throughout the southern Appalachian basin indicating temperatures of about 1750C that lasted less than 1 m.y. This temperature is higher than the regional CAI (about 80~ Roden et al (1993) suggest that fluid flow was apparently local- ized in or near the K-bentonites, partially annealing zircon fission tracks, completely annealing apatite fission tracks, but not affecting conodont alteration indices on a regional scale.

A N A L Y T I C A L METHODS

Sampling

Figure 1 shows the locations of the 56 K-bentonites used in this study (their approximate geographic coordinates are included in Table 1). Eight K-bentonites

were collected from core (samples: 11-13 and 57-60) and the rest collected from outcrops. Samples 6, 22, 23, 25 were collected by the authors and all others were obtained from colleagues.

Random powder preparation

Samples were lightly crushed in an Alundum mortar and disaggregated in distill sd water with an ultrasonic probe. The < 1.0-~tm (equiNalent spherical diameter) and/or < 0.5-/~m size fractiOns were separated by t imed centrifugation. Clay suspensions were saturated with 1 M CaCI2 to induce flocculation to a id in collection and to exchange interlayer cations with a common ion in order to mainta in comparable expandabil i ty measure- ments between the samples. All samples were subse- quently washed repeatedly to remove excess chloride until the discarded wash showed no evidence of chloride when tested with AgNO3. Washed samples were concentrated by ultracentrifugation. The core samples were obtained as clay separates and their preparat ion is described by Elliott and Aronson (1987).

All clay separates were redispersed in 50 ml of dis- tilled water for freeze-drying. The suspensions were frozen in a Labconco shell freezer at - 3 0 ~ and ly- ophil ized in a Labconco 4.5 freeze drier. The freeze- dr ied clay was ground lightly with a rubber s topper mounted on a glass rod. Immedia te ly pr ior to X R D analysis, the freeze-dried outcrop separates were dehydrated in an oven at either 250~ or 350~ for one h. The core sample separations were spread in an evap- orating dish and solvated with ethylene glycol vapor for 24 h at 60~ before preparing a random powder X-ray diffraction (XRD) mount. The core separates were treated with ethylene glycol solvation instead of heating because they were to be used in a future study for which heating was inappropriate. All randomly oriented sample mounts were prepared for X R D by a side- loading method using a sample holder 3 cm long, 2.35 cm wide, and 0.15 cm deep that accommodates approximately 600 mg of clay.

Orientated aggregate preparation

Oriented X R D preparat ions were made by dispers- ing approximately 200 mg of clay separate in 2 ml of water, pipetting the suspension onto a glass slide, and drying at room temperature. Oriented X R D mounts were analyzed air-dried and after vapor solvation with ethylene glycol at 60~ for 12 h.

X-ray diffraction analysis

Diffraction data were collected with a Siemans D-500 diffractometer equipped with a digital step counter, a diffracted beam graphite monochromator , 1 ~ beam and 0.15 ~ detector slits and using CuKa radiation. All step scans were conducted at 20 increments o f 0.05 ~ Count t imes for the random powder preparat ions were 20 s per step and runs were made from 16--44 ~ Runs for

274 McCar ty and Reynolds Clays and Clay Minerals

Table 1. Sample locat ions and structural parameters .

Decimal Decimal Bentonite bed long. lat. Sample # and l o c a t i o n E x p . % R e i c . P o P ~ Peo

- 86.00 34.00 2 Deicke AL 11 3 50 25 70 - 8 6 . 0 0 34.00 3 Deicke A L 18 1.5 75 30 85 - 85.00 38.20 4 Deicke KY 21 1 50 80 60 - 92.70 44.42 5 Deicke M N 22 1 45 60 60 - 9 0 . 5 3 38.25 6 Deicke M O 29 1 60 60 50 - 78.50 40.58 7 Ordovic ian PA 11 3 70 82 30 - 78.00 39.45 8 Deicke W .VA 11 3 80 60 30 - 8 4 . 8 5 38.00 9 Ordovic ian KY 18 1.5 70 75 20 - 8 6 . 1 7 39.83 10 Ordovic ian IN 21 1 65 60 25 - 8 5 . 2 0 40.75 11 Ordovic ian IN 16 3 70 75 35 - 85.08 41.17 12 Ordovic ian IN 13 3 55 60 60 - 85.08 41.17 13 Ordovic ian IN 22 1.5 60 85 20 - 8 4 . 8 0 38.30 14 Ordovic ian KY 24 1 70 70 40 - 73.83 43.13 15 Ordovic ian NY 18 1 60 65 40 - 85.45 34.92 16 Millbrig G A 14 3 85 55 60 - 8 5 . 1 0 34.20 17 Deicke G A 9 3 65 40 90 - 9 0 . 5 3 38.25 18 Deicke M O 26 1 50 50 50 - 90.63 38.53 19 Millbrig M O 27 1 50 50 50 - 8 4 . 7 2 38.93 20 Deicke KY 29 I 55 85 60 - 84.67 39.12 21 Ordovic ian K Y 15 2.5 75 50 75 - 9 0 . 5 3 38.25 22 Millbrig M O 27 1 60 65 65 - 9 0 . 5 3 38.25 23 Millbrig M O 27 1 60 60 45 - 8 4 . 5 8 37.83 24 Millbrig K Y 26 1 60 60 65 - 9 0 . 5 3 38.25 25 Deicke M O 19 1 50 60 60 - 9 1 . 3 3 42.87 26 Nasse t Ord. IA 15 3 75 35 95 - 76.00 43.00 27 Ordovic ian NY 18 1.5 62 75 45 - 7 7 . 5 0 40.90 28 Ordovic ian PA 14 2.5 75 60 40 - 77.25 41.25 31 Ordovic ian PA 12 3 60 80 25 - 71.25 46.75 32 Ordovic ian Quebec 26 1 50 90 20 - 84.53 38.15 33 Ordovic ian KY 15 3 45 55 60 - 76.17 42.97 34 Tioga N Y 10 3 60 80 35 - 8 3 . 4 7 36.57 35 Deicke T N 10 3 50 30 60 - 8 5 . 9 2 36.25 36 Deicke T N 30 1 65 75 46 - 8 1 . 0 8 37.25 37 Deicke VA 9 3 50 30 80 - 8 3 . 0 0 36.70 38 Deicke VA 13 2.5 85 35 90 - 9 2 . 7 0 44.42 39 Deicke M N 26 1 60 60 45 - 9 1 . 6 0 43.58 40 Millbrig M N 27 1 30 90 35 - 8 5 . 9 2 33.70 41 Millbrig A L 11 3 70 55 60 - 74.42 42.88 42 Ordovic ian N Y 19 2.5 55 85 30 - 75.17 43.17 43 Ordovic ian NY 11 3 45 65 45 - 7 4 . 7 0 42.78 44 Tioga N Y 13 3 55 75 15 - 75.92 42.92 45 Tioga N Y 13 3 50 98 20 - 7 6 . 7 8 42.85 46 Tioga NY 9 3 50 85 10 - 7 7 . 2 7 42.93 47 Tioga N Y 12 3 55 97 15 - 78.73 42.88 48 Tioga N Y 21 1.5 55 80 45 - 73.83 43.13 49 Ordovic ian N Y 18 1.5 55 85 10 - 91.00 38.72 50 Millbrig M O 33 1 50 50 50 - 90.83 38.35 51 Deicke M O 29 1 50 50 65 - 9 0 . 0 8 37.92 52 Deicke M O 31 1 60 50 50 - 8 9 . 2 0 43.33 53 Ordovic ian WI 19 1.5 65 45 75 - 8 4 . 5 0 39.08 54 Deicke K Y 15 2.5 50 90 45 - 7 5 . 6 7 45.50 55 Ordovic ian Ot tawa 13 2.5 70 60 40 - 75.67 45.50 56 Ordovic ian Ot tawa 21 2.5 50 95 35 - 8 6 . 5 0 38.75 57 Ordovic ian IN 24 1 60 95 30 - 8 6 . 5 0 38.92 58 Ordovic ian IN 22 1 70 80 25 - 85.13 38.47 59 Ordovic ian K Y 26 1 60 90 30 - 86.83 39.33 60 Ordovic ian IN 19 1.5 45 92 40 Min. 9 1 30 25 l0 Max. 33 3 85 98 95 Average 19 2 59 66 47

Vol. 43, No. 3, 1 9 9 5 Rotationally disordered illite/smectite in Paleozoic K-bentonites 275

the oriented preparations were made at 1 to 5 s per step from 2-50 020. These conditions produced a peak height intensity of approximately 10,000 counts for the mica prism reflection near 19.9 *20 (the 020;110).

I/S expandability and Reichweite determinations were made from the oriented aggregate diffraction patterns using the delta-20 method described in Moore and Reynolds (1989) and in some cases by means of the NEWMOD computer program (Reynolds 1985).

Computer modeling o f three-dimensional structures

The three-dimensional structural parameters of percent cv, percent tv, and the percent and type of rotational disorder in the I/S were determined by modeling each diffraction pattern with the computer program WILDFIRE (Reynolds 1993). No attempt is made here to describe the WILDFIRE algorithm and the reader is referred to Reynolds (1993) for a complete descrip- tion.

All of the calculated diffraction profiles are based on atomic coordinates and unit cell parameters for cv and tv unit cells from Tsipursky and Drits (1984) and Drits et al (1984). The cv unit cell dimensions are: a = 5.199 A, b = 9.005 ~ , c = 10.09/~ and 3 = 99.13". The tv unit cell dimensions are: a = 5.199/~, b = 9.005/k, c = 10.164/~ and 3 = 101.3*. In every case the calculated difliaction patterns were based on crystals containing 60 uni t cells along X and 30 unit cells along Y. Crystal size in the Z direction was determined on a case by case basis as part of the WILDFIRE modeling proce- dure. The number of unit cells in the Z direction de- pends on a particle size distribution that is calculated from I/S expandability and Reichweite (see also Al- taner and Bethke 1988). Particle size distributions used in the calculated diffraction patterns that are shown later as examples are presented in Table 2. Fe and K were not varied in any of the calculated diffraction patterns from the default values of 0.14 and 0.7 atom per formula unit, respectively. Composition, atomic coordinates, and crystal dimensions in X and Y were not varied in WILDFIRE, and the particle thickness distribution was estimated from basal diffraction data. Therefore, all of the diffraction patterns in this study were modeled with combinations of just three adjust- able input parameters Pc~, Po, and P6o. Pcv refers to the percentage of cv layers interstratified with tv layers in the calculated model crystallite structure where:

P , + P,v = 100.

Po is the percentage of layers rotated zero degrees with respect to a preceding adjacent layer. P0 = 100% describes the 1Mpolytype where all layers have the same orientation. For structures with rotational disorder of type n. 120 ~ P,2o- = P24o.. IfPo = 33% there are equal proportions of interlayers involving 0 ~ 120 ~ and 240 ~ rotations of adjacent layers. Po = 33% produces the

Table 2. Crystallite thickness distribution for calculated patterns shown in Figures 5-10.

Sample

34 8 45 26 18 40 Tioga Deicke Tioga Nasset Dcicke Millbrig

N' NY W.VA NY IA MO MN

1 0 0 0 0 0 0 2 0.067 0.067 0.067 0.272 0.512 0.422 3 0.188 0:188 0.188 0.228 0.269 0.268 4 0.285 0.285 0.285 0.217 0.129 0.161 5 0.193 0.193 0.193 0.140 0.55 0.081 6 0.129 0.129 0.129 0.88 0.023 0.04 7 0.084 0.084 0.084 0.055 8 0.054 0.054 0.054

N t is the number of 2:1 layers per crystallite.

greatest amount of disorder or highest degree o f " l M d - hess" possible. If P, < 33% then the structure is partially ordered with respect to the 120* and 240 ~ rotations, as are 2M1 and 3 T structures. This latter type of nearest-neighbor or next-nearest neighbor ordering does not pertain to any of the samples or modeled structures in this study.

Rotations of n- 60 ~ (n = odd integer) are possible and are not crystallographically equivalent to rotations of n- 120 ~ but rotations of 60", 180 ~ and 300 ~ are crys- tall,graphically equivalent to each other. Pto refers to the proportion of layers that are rotated n. 60* out of the total fraction of rotated layers (P0 not being considered a rotated layer). If /6o = 0 all of the rotational disorder is of the n. 120 ~ type. IfPto = 100% rotations of 120* and 240 ~ are eliminated, producing a random stacking sequence of 60 ~ 180 ~ and 300 ~ rotations. When Pto = 60% there are equal proportions of 60 ~ 120 ~ 180 ~ 240 ~ and 300 ~ rotations.

Intensities of diffraction peaks from the experimental random powder diffraction patterns are affected by preferred orientation of the clay powder in the sample holder. The calculated patterns are compensated for this effect by means of the March function correction factor (Dolasse 1986). A correction factor of 1.0 describes perfect random orientation. Correction factors of less than 1.0 simulate the effects of preferred orientation where proportionally more crystals in the aggregate mount are oriented with d(001) parallel to the surface of the mount. A correction factor was obtained by measuring departures of the 003/(020,110) intensity ratios from the values calculated for random orienta- tions.

WILDFIRE calculates basal reflections separately and adds them to the three-dimensional calculated reflections after normalizing for absolute intensity and preferred orientation. Special attention was paid to modeling the basal reflection part of the composite three- dimensional calculations by first modeling the basal part of the measured random diffraction patterns with NEWMOD (Reynolds 1985) and then using the de-


002

002

020 ; t 110

h ] 003; 022

112

003; 022

200 ; ]31 30 0\ A

112

12 h3

0 % Cis-Vacant

201; 132

0~0; 133202;

100 % Cis-Vacant 201;

~ 32 202; 040; 133

t__, it8 , i , L , i , L , i , ~ , i , i . , L , I , i , ~ , 20 22 24 26 28 0 32 34 36 38 40 42

~

Figure 3. Calculated diffraction patterns for 0% cv and 100% cv structures. The parameters are: Po = 70% and all rotated layers are n- 120" (P~2o = 15%;/'2,o = 15%; P6o = 0) in both calculations.

termined 001 parameters in the WILDFIRE calculations. The basal reflections contain no information about three-dimensional disorder and the octahedral cation distributions. They are included here to model correctly interferences between three-dimensional and basal reflections.

Diffraction pattern interpretation

The diagnostic hkl reflections used to determine Pcv and Po, and the unit-cell type of a mica polytype are o f the category k ~ 3n. These peaks, and all hkl reflections, are broadened by small crystal size and/or rotational disorder. To extract model Pcv and Po values requires an analysis o f the 02l; 111 reflections between 19 and 34 020. The region between 34 and 39 ~ consists of the k = 3n reflections 20l; 131 and is useful for determining the quantity of n. 60 ~ rotations,/60- Both n. 60* and n. 120 ~ rotations have very similar effects on the 02l; 111 peaks (Drits et al 1984, Sakharov et al 1990, Drits and Tchoubar 1990, Reynolds 1993, this study). Modeling o f experimental diffraction patterns with W I L D F I R E consists of adjusting the input parameters Po, Pcv, and/ '60 until agreements of peak position (d-value) and peak shape are optimized between the experimental and the calculated diffraction patterns.

Figure 3 illustrates the differences between a disordered tv I/S (1Md) and a disordered cv I/S (1Md). Both o f these calculated patterns have Po = 70% and the rotations o f the remaining 30% of layers are of the n. 120 ~ type. The hk/indices of diagnostic peaks are shown

PO = 7 0 %,Pcv=50%

a60 ( % )

100

80

60

40

20

i 316 34 38

~ Figure 4. Calculated diffraction patterns demonstrating the effects of n- 60 ~ rotational disorder on the 20l; 13l reflections. The parameters are: Po = 70%, and Pcv = 50% for all of the calculations.

on the figure; Reynolds (1993) discussed the details of the peak positions and relative intensities.

Figure 4 illustrates the effect of increasing values of P60 on the 20l; 131 reflections. The calculations in Fig- ure 4 were all made with Po = 70%, Pcv = 50%, and the P6o values shown on the figure. P6o values are the fraction of the 30% of rotated layers that consist of equal proportions of 60 ~ 180 ~ or 300 ~ rotations. These calculations show the migration o f the 202; 131 and 201; 132 reflections toward each other so that when all o f the rotated layers are o f the n. 60 type there is a single broad peak at about 37 ~

V o l . 4 3 , N o . 3 , 1 9 9 5 R o t a t i o n a l l y d i s o r d e r e d i l l i t e / s m e c t i t e i n P a l e o z o i c K - b e n t o n i t e s 2 7 7

e~

e~

0

0 0 0 0 0 0 0 0 0 0 0 o o o o 0 0 0 0 0 0

I I ~ I I I I I I I I ~ 1 I I I

~ 0 0 0 0 0 0 ~ 0 0 0 0 ~ 0 ~ 0 0 0 0 0 0

I I I I I I I I I I I I I I I I I

~ I I I I I I

~ I I ~ ~ ~ 1 ~ - -

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

I I I I I I I I I I I I I I I I I

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

I I I I I I I I I I I I I I I I I I I I

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

~ . ~ . ~ . ~ . . . . . . . . . . . ~

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

I I I I I I I I I I I I I I I I I I I I

. . . . . m o ~ m ~ m ~ ~ m ~

O 0

278 McCarty and Reynolds

Table 4. Structural formulae.

Clays and Clay M~nerals

Inter- Inter- Oct. Inter- layer layer Oct. Oct. Oct. cation Tet. Oct. 2:1 layer

K Ca Tet. Si Tet. A1 AI Fe Mg sum charge charge charge charge Sample Units atoms atoms atoms atoms atoms atoms atoms atoms

41 Millbrig AL 0.57 0.12 3.57 0.43 1.48 0.16 0.35 1.99 -0.43 -0.38 -0.81 0.81 55 Ordov. Ottawa 0.67 0.15 3.37 0.63 1.52 0.08 0.42 2.02 -0.63 -0.34 -0.97 0.97 56 Ordov. Ottawa 0.49 0.12 3.27 0.73 1.71 0.12 0.26 2.09 -0.73 0.03 -0.7 0.73 42 Ordov. NY 0.54 0.08 3.2 0.8 1.68 0.2 0.23 2 . 1 1 -0.8 0.1 -0.71 0.70

70rdov. PA 0.61 0.06 3.26 0.74 1.6 0.16 0.36 2.12 -0.74 -0.003 -0.74 0.73 8 Deicke W.VA 0.59 0.08 3 . 2 1 0.79 1.74 0.1 0.26 2.1 -0.79 0.05 -0.74 0.75

49 Ordov. NY 0.54 0.09 3.26 0.74 1.56 0.25 0.3 2.11 -0.74 0.03 -0.71 0.72 16 Millbrig GA 0.63 0.08 3.35 0.65 1.31 0.36 0.44 2 . 1 1 -0.65 -0.13 -0.78 0.79 17 Deicke GA 0.51 0 . 1 1 3.39 0.61 1.35 0.37 0.36 2.08 -0.61 -0.13 -0.74 0.73 18 Deieke MO 0.56 0.16 3.34 0.66 1.35 0.29 0.42 2.06 -0.66 -0.22 -0.88 0.88 40 Millbrig MN 0.6 0.14 3.39 0.61 1.45 0.2 0.39 2.04 -0.61 -0.27 -0.89 0.88 27 Ordov. NY 0.49 0.07 3.3 0.7 1.69 0.17 0.24 2.1 -0.7 0.07 -0.63 0.63 34 Tioga NY 0.7 0.07 3 . 4 1 0.59 1.49 0.13 0.44 2.06 -0.59 -0.24 -0.83 0.84 39 Deicke MN 0.54 0.15 3.43 0.57 1.51 0.15 0.37 2.03 -0.57 -0.27 -0.84 0.84

3 Deicke AL 0.65 0.06 3.44 0.56 1.27 0.29 0.55 2 . 1 1 -0.56 -0.21 -0.77 0.77 26 Nasset IA 0.67 0.14 3 . 5 1 0.49 1.29 0.13 0.64 2.06 -0.49 -0.46 -0.95 0.95 24 Millbrig KY 0.58 0.16 3.5 0.5 1.42 0.13 0.48 2.03 -0.5 -0.41 -0.91 0.90 19 Millbrig MO 0.59 0.2 3.42 0.58 1.44 0.13 0.43 2 -0.58 -0.43 - l .01 0.99 53 Ordov. WI 0.62 0.27 3.39 0.61 1.23 0.25 0.51 1.99 -0.61 -0.54 -1.16 1.16 45 Tioga NY 0.69 0.25 3.33 0.67 1.47 0.1 0.39 1.96 -0.67 -0.51 -1.18 1.19 21 Ordov. KY 0.57 0.31 3.5 0.5 1.28 0.14 0.52 1.94 -0.5 -0.69 - l . 19 1.19

Chemical analysis

Bulk chemical analyses were made by X-ray fluo- rescence methods (XRF) on 21 clay-fraction separations (Table 3) by X-ray Assay Laboratories, Don Mills, Ontario. These samples were chosen on the basis o f 1) structural diversity after X R D and WILDFIRE modeling, 2) geographic distribution, and 3) adequate sample size (1.5 g). Each separate sent for X R F analysis was checked for contamination by random-powder X R D analysis.

RESULTS

Chemical composition

Table 3 shows chemical analyses and Table 4 shows calculated US structural formulae based on an O~o(OH)2 formula unit. Total Fe was calculated as Fe 3§ . Samples 19, 53, 45 and 21 all have anomalously high 2:1 layer charges. This high layer charge may be due to small amounts o f secondary mineral phase(s), based on closer inspection o f the random powder diffractograms from the separates submitted for analysis. These samples were not used in subsequent structure-composition correlation.

X-ray diffraction and modeling

Table 1 lists sample name and location (Figure 1), along with the I/S structural parameters of percent expandability, Reichweite determined from oriented- sample X R D and N E W M O D modeling, and the parameters P0, Pcv, and/)6o determined from modeling

each random diffraction pattern with WILDFIRE. There is greater uncertainty in accuracy of the/)60 parameter than there is for Po and Pcv, because the 20l; 131 diffraction profile used to model/)60 is also affected by crystal shape, Fe content, and tetrahedral tilt and rotation (Reynolds 1993).

Pcv is less accurate ifPo is low. IfPo is near the lower limit of 33% the diagnostic k ~ 3n reflections are so broad that a large change in the Pcv input parameter has little or no visible change in the calculated diffraction pattern. When P0 values increase and the k ~ 3n reflections become better resolved, the accuracy of the corresponding P , values increases. The nature of rotational stacking defects and their effect on peak position and the accuracy of cv determination are examined in detail by Drits and McCarty (1995).

Figures 5-10 show random-powder diffraction patterns o f dehydrated I/S along with their corresponding WILDFIRE calculated patterns. They are presented as examples that show a range o f the modeled Po, Pcv, and P6o parameters and are representative o f all of the samples and corresponding calculated patterns. The com- parisons demonstrate the utility of calculated patterns in realistically modeling three-dimensional I/S crystal structures involving interstratification o f tv and cv unit cells combined with different types and amounts o f rotational disorder.

Dashed lines in Figures 5-10 are used to show the 20 coincidence of the diagnostic peak positions men- tioned above and shown in Figure 3. Correspondence with peak shape was made by visual inspection and achieved by trial and error adjustment o f P0, Pcv, and

020;110

112 ~13 I l l ~12 ~ '2 ; 131 201;

, , I32

020; 1 I0

P@ = 60%

P = 80% tF

i P60 = 35%

II �9 .~m

~ Cu K a radiation

Figure 5. Comparison between calculated (top) and experimental diffraction patterns for Tioga sample 34. The parameters are: Po = 60%, Pcv = 80% and P6o = 35%.

P6o- Comparison of the d-values for the calculated and experimental k 4: 3n reflections for the diffraction patterns in Figures 5-10 is shown in Table 5.

Figure 5 shows data for the Devonian Tioga K-bentonite, in central New York near Syracuse enclosed in the Onondaga Limestone Formation. This is an example of a good fit between a calculated and experimental diffraction pattern. The I/S has 80% cv interstratification, and 60% of the layers have 0 ~ rotations. Pto is relatively low at 35%, indicating that the structure is dominated by n. 120 ~ rotations, as evidenced by the well-modulated peaks between 36 and 39 ~ This Tioga sample is a "classic" example of a 1 M d cis-

vacant I/S. Sample 8 shown in Figure 6 is from the middle Or-

dovician Deicke K-bentonite near Martinsburg, West Virginia. This Deicke sample is a cv-dominated I/S comparable to the Tioga sample in Figure 5, except it is more ordered as seen by the sharper more intense i 12 and 112 peaks, which are also more intense because the structure has fewer cv layers. The diffraction pattern was modeled with Po = 80%, compared with Po = 60% for the Tioga sample, and Pcv is 60% in this sample compared with 80% for the Tioga. Like the Tioga US, this sample is dominated by n. 120 ~ rotations (P6o = 30%).

Sample 45 shown in Figure 7 is from the Tioga K-bentonite at Jamesville, New York. It demonstrates the diffraction effects produced from a very high proportion of cv layers combined with a high degree of rotational disorder indicated by a Po value of only 50%. The i 12 and 112 reflections which are prominent in the tv structure are now just broad shoulders on both sides of the 003 peak; the 111 and i 13 peaks are broad,

Vol. 43, No. 3, 1 9 9 5 Rotationally disordered illite/smectite in Paleozoic K-bentonites 279

112 202; 201;

i 13 023 131 732

A

P0 = 80%

pc, = 6 0 %

P6 O = 30%

__De. i.ck r <0.5-~m

I , 118 , 20 ' 212 . 214 . 216 ' 218 30 312 34 ' 316 318 ' 4~ '412 ' '

~ Cu KI~ radiation

Figure 6. Comparison between calculated (top) and experimental diffraction patterns for Deicke sample 8. The parameters are: Po = 80%, Pc~ = 60% and Pto = 30%.

but well defined. The rotational disorder is nearly all n- 120 ~ indicated by a P60 value of only 20%.

Sample 26 shown in Figure 8 is from the Nasset K-bentonite at Guttenberg, Iowa. This US is relatively well ordered with P0 = 75%. The low Pcv value of 35% eliminates the 111 peak and the T 13 is a broad hump; at the same time the definition and intensity of both the i 12 and 112 peaks are enhanced. Rotations are almost all n.60" as shown by the single broad peak near 37 ~

Sample 18 shown in Figure 9 is from the Deicke K-bentonite in the southern Illinois basin and is very characteristic of other samples from that area. The combinat ion ofPo, Pcv, and Pto values all equal to 50%

020;110

I I I 713 131 732

Po = 50%

Per = 98%

P60 = 20%

Tioga ~0.S-~m NY

i , i , , i i , i , i , i i , i , i , i , 412 , t 18 20 22 24 26 2g 30 32 34 36 38 40

~ C u K0t radiation

Figure 7. Comparison between calculated (top) and experimental diffraction patterns for Tioga sample 45. The parameters are: Po = 50%, P , = 98% and/'6o = 20%.

280

020; 110

McCarty and Reynolds

020; 110

Clays and Clay Minerals

/11 ~12 023 ~ 20/: 131 112 TI3

Po = 75% ~v= 35%

P60 = 95%

~ I ~ 'l ', ~ / / ! V ~! ~ , ~ ? / ~ N~s~et <l.O-~tm

, + 2~ 2'~ 2, 2'~ 2'+ 3'0 3'2 3',';,'3'+ ~ 4 2 + ~ Ca K~ radiation

Figure 8. Comparison between calculated (top) and experimental diffraction patterns for Nasset sample 26. The parameters are: Po = 75%, Pcv = 35% and/ '6o = 95%.

p r o d u c e s a u n i q u e d i f f rac t ion p a t t e r n in w h i c h the i 12 p e a k is b r o a d , b u t re la t ive ly well def ined, a n d the 112 a n d ] 13 peaks b e t w e e n 28 a n d 32 *20 h a v e m e r g e d i n t o a b r o a d s lop ing shoulder . T h e re la t ive ly h igh per- cen tage o f n- 60* r o t a t i o n s p roduces a single peak nea r 37 *20.

S a m p l e 40 s h o w n in F igure 10 is f r om the Mi l lb r ig K - b e n t o n i t e in s o u t h e a s t M i n n e s o t a . Po is 33% ind i - ca t ing m a x i m u m r o t a t i o n a l d i sorder , a n d all o f the d i agnos t i c peaks are b r o a d e n e d so m u c h t h a t they are undef ined . T h i s is a n e x a m p l e o f a n e n d - m e m b e r 1Md s t ruc ture . Large changes in Pcv h a v e l i t t le v i s ib le effect

020;[10

~12 I I 201; 131

PO = 50% Pc ~ = 50% P60= 50%

, ' ~ 20'2+2 24 26 ~+ ;0 3 2 34'36 3~+ ~ , ' 2 +20 Cu Ko~ radiation

Figure 9. Comparison between calculated (top) and experimental diffraction patterns for Deicke sample 18. The parameters are: Po = 50%, Pcv = 50% and P6o = 50%.

7J01; 131

I ~ PO = 3 3 =90%

: P60 = 35%

I i . -ttm

(8 20 212 24 ~ 6 218 30 32 314 316 318 40 42 ~ Ca KCx radiation

Figure 10. Comparison between calculated (top) and experimental diffraction patterns for Millbrig sample 40. The parameters are: P0 = 30%, P+v = 90% and P6o = 35%.

o n t he ca lcu la ted d i f f rac t ion p a t t e r n a n d the re fo re the accuracy o f the P+v va lue is low.

These examples d e m o n s t r a t e the wide var iabi l i ty t ha t exis ts w i t h i n the genera l class o f d i so rde red I/S, a n d the p o w e r W I L D F I R E has in accura te ly m o d e l i n g I /S t h r e e - d i m e n s i o n a l crystal s tructures. In some cases, such as F igure 8, the 002 a n d 003 peaks in the e x p e r i m e n t a l d i f f rac t ion p a t t e r n i nd ica t e a lower d -va lue t h a n the c o r r e s p o n d i n g peaks in the ca lcu la ted d i f f rac t ion pa t - tern . T h i s d i s c r epancy suggests t h a t the defau l t v a l u e o f d(001) = 9 . 7 / ~ used for d e h y d r a t e d smec t i t e m a y n o t h a v e b e e n app rop r i a t e , b u t th i s q u a n t i t y ha s n o

Table 5.

d-value d-value experimental calculated

Sample hid (,~) (A)

34, Tioga 111 3.87 3.87 NY 112 3.60 3.59

112 3.12 3.12 113 2.89 2.87

8, Deicke 111 3.88 3.88 W.VA 112 3.64 3.62

112 3.09 3.10 T13 2.90 2.88 023 2.69 2.67

45, Tioga 111 3.85 3.87 NY 113 2.87 2.87

26, Nasset 111 4.38 4.34 IA 112 3.65 3.63

112 3.09 3.09 113 2.91 2.90

18, Deicke 112 3.65 3.63 MO

40, Millbrig -- -- -- MN

Voi. 43, No. 3, 1995 Rotationally disordered illite/smectite in Paleozoic K-bentonites 281

k 00 o ~ r = 0 . 7 8 0

0

75 ;) o o

o o o ..: + o o O ; . . : :

o o o

o o

0 I I I 20 40 Bo 8 0

I OO

o 80

oo

60

o o

4o

0 20

r = 0.278

o o

o o o o o o c o

o o o o o o o o o o

o o o o o o o o o o o o

o o o o

o I I I

40 60 8o 100

[] []

Figure 11. Structural relationships from all of the I/S sam- pies modeled in this study: (1 la) P~ vs P6o; (1 lb) P~, vs Po.

effect on the positions of the diagnostic hkl peaks or the estimation of Pcv, Po and P+o-

DISCUSSION

Structure-composition relationships

Three-dimensional structural parameters. Figure 1 la shows an inverse relationship between the proportion of cv interstratification and the proportion o f n-60* rotations in the rotated layers. In Figure 1 lb, Pcv is plotted against Po and shows no apparent relationship between the proportion of cv interstratification and the proportion o f layers with a 0 ~ stacking angle. The parameters Pcv, P0, and/'6o do not show any systematic correlation with US expandability and Reichweite.

Relationships between composition and the proportion of interstratified cis-vacant layers. Figure 12 shows a significant linear relationship between composition and cv interstratification (P~) for the I/S samples for which we have both chemical data and structural data. A decrease in cv interstratification correlates with an increase in Mg and Fe in I/S 2:1 layers where Mg and Fe substitute for Al in the octahedral sheets.

Layer charge relationships. Figure 13a shows a linear correlation between the increases in magnitude of both calculated octahedral charge and total 2:1 layer charge

Figure 12. and P,.

r = 0.688

100 - o

BO - oo

60 -

,*o

O

20 I I I I I

4 5 6 7 8

F e 2 0 8 + M g O ( w t %)

Relationship between composition (Fe203 + MgO)

-07 /

o o o

o 8

e~ 09

o -I I I i

-0.4 -0 2 0 0.2 ] O c t a h e d r a l C h a r g e

-0.7

O9

-1 -o9

[]

0 r = 0 . 4 8 5

O �9 o o o

o o

o ~

o o

I I ~ t

-08 -0.7 -0.6 -0.5 -o4 T e t r a h e d r a l C h a r g e

-035

-O.4

0.45

05

"~ -055

-0.6

0"650.4

[]

Srodoh et al., 1986 P o l a n d

r ~ 0.309 Q

o O

o o O o 0

o

~ I I I

035 -03 0.25 -0.2

O c t a h e d r a l C h a r g e

Srodoh el al., 1986 Poland -o35

-0.45 -0.5 -04 r =0894 ~ ~ O ~ O ~ O O O

-055

-0.6

-0.65 I I I I I -035 -0.3 -0.25 -02 -0.l 5 -0.1 -0.05

[ ] T e t r a h e d r a l C h a r g e

o.6

0.65

-0.7

~ -o.75

SrodohetaL, 1986 U.K.

O r=0.14

0 0

o

o o

o -0.8 I I I

-0.4 -0.35 -0.3 -0.25 - 0 2

[] Oetahedral Charge

S r o d ~ et aL, 1986 U . K . -0.55

r = 0.774 O

-0.6

-065 /

-O.7

-0.73

O -O.g i , i i I

-0.5 -0.45 -0.4 -0.35 -0.3 -0.25 -0.2

[ ] T e t r a h e d r a l C h a r g e

Figure 13. I/S charge relationships, calculated octahedral and tetrahedral charge vs calculated 2:1 layer charge. Figures (13a-b) this study. Figures (13c--d) are from Srodofi et al (1986), a suite of I/S samples from an Upper Carboniferous bentonite Upper Silesia, Poland. Figures (13c--d) are from Srodofi et al (1986), a suite of I/S samples from Silurian bentonites in the United Kingdom.

for the I/S samples analyzed in this study. Figure 13b shows no such systematic correlation between calculated tetrahedral charge and calculated 2:1 layer charge.

The finding that 2:1 layer charge does not have a linear relationship with tetrahedral charge is in contrast to published data for other suites o f I/S samples from bentonites. Figure 13c shows octahedral charge vs 2:1 layer charge, and Figure 13d shows tetrahedral charge vs 2:1 layer charge in I/S samples from Upper Car- boniferous bentonites Upper Silesia, Poland (Srodofi et al 1986). In these samples tetrahedral charge pro- gressively increases with 2:1 layer charge whereas no such relationship exists for octahedral and 2:1 layer charges.

Figures 13e and 13f also show no correlation be-


Figure 14. Side view projection of basic mica structure showing direction of shortening from shared octahedral edges and increasing octahedral sheet lateral dimensions from octahedral Mg substitution.

Figure 15. Basal oxygen positions in octahedral and prismatic coordination.

tween octahedral and 2:1 layer charge, but show a pro- gressive increase in both tetrahedral charge and 2:1 layer charge in I/S samples from Silurian bentonites in the United Kingdom (Srodofi et al 1986).

No information is available on the three-dimensional structure for the Polish or United Kingdom sam- pies but the data suggest that the crystal chemistry of I/S may be quite different in separate geologic locations.

Crystal chemistry

In this suite of I/S samples from K-bentonites, n. 60 ~ rotations are correlated with samples that have the highest content of octahedral Mg and Fe and tetrahedral Si. This observation can be reconciled by crystal chemical considerations of the basic mica structure. Tetrahedral thickening, tilt, rotation and thinning of the octahedral sheet are four mechanisms of adjustment for the misfit between the tetrahedral and octahedral sheets in the 2:1 layer. The nature and extent of these adjustments have different consequences for the subsequent rotational stacking angles in the crystal (Radoslovich 1959, Radoslovich and Norrish 1962, Bailey 1966, 1984). Shortening of shared octahedral edges in the direction diagonal to the sheet surface occurs as octahedral anions move toward each other to shield octahedral cation repulsion (Bailey 1984; Fig- ure 14). Therefore, the entire octahedral sheet is thinned and expanded along X and Y. Lateral octahedral edges around the vacant octahedral sites are expanded as a result of this shared-edge shortening (Drits et al 1993, Bailey 1984). Substitution of Mg for Al in the octahedral sheet increases the lateral dimensions of the sheet. The tetrahedral sheet b-axis dimension is minimized with more Si and less A1 substitution. There- fore, octahedral Mg substitution and lower tetrahedral A1 reduces the misfit for a given overall 2:1 layer charge. This type of layer composition minimizes tetrahedral tilt and rotation, making a more regular structure. When

misfit reduction is accomplished by tetrahedral tilting the bridging basal oxygen between adjacent tetrahedra will be depressed below the other two basal oxygens of each tetrahedron (Bailey 1984). The depressed oxygens form grooves or corrugations in the plane of basal oxygens (Bailey 1984).

When tetrahedral tilt and rotation are minimized the basal sheets can more easily coordinate with prismatic 6-fold coordination (Figure 15) around the K ion where the oxygen atoms in adjacent layers are superimposed over one another (the n. 60 ~ rotation type). This type of interlayer cation bonding differs from 6-fold octahedral coordination (Figure 15) where the oxygen atoms are staggered in adjacent layers (the n. 120 ~ rotation type). Lower Al content in the tetrahedral sheet reduces the amount of tetrahedral tilt and rotation needed for tetrahedral sheet and octahedral sheet ar- ticulation and also results in less basal oxygen corrugation. With less corrugation the basal oxygens are in a common plane, making n. 60 ~ and n. 120 ~ rotations more electrostatically equivalent. When tetrahedral Al substitution is high and octahedral Mg substitution is low, n. 120 ~ rotations are the most favored because that is the best configuration to minimize basal oxygen repulsion.

Prismatic coordination does occur, although not as frequently, with Al-rich tetrahedra and octahedra in 2 M 2 muscovite. Muscovite with the 2M2 structure differs from the I/S described in this study in both composition and layer stacking. The 2M2 structure consists of regular alternations of layers by -+ 60 ~ and a consis- tent 1:4 A1 to Si tetrahedral composition, whereas the I/S has a variable composition and a 1M stacking sequence with random rotations of n- 60 ~ More work is needed to fully understand the structural relationships and physical conditions that result in the 2M2 structure. Another example of prismatic coordination is docu- mented in glauconite by Sakharov et a! (1990). An average glauconite composition is similar, although not identical, to the I/S samples from K-bentonites (Bailey

Vol. 43, No. 3, 1 9 9 5 RotationaUy disordered illite/smectite in Paleozoic K-bentonites 283

1980; Table 4). Like the US, glauconite has significant Mg and Fe octahedral substitution, but these two minerals formed in vastly different environments. It is not known what relative influences crystal-chemistry and environmental formation conditions play in determining the structural features and rotational stacking defects in I/S and other 2:1 layer phyllosilicates. In a hydrothermal alteration zone (Saskatchewan, Canada), Drits et a l (1993) found illites with different octahedral cation distributions between t rans and c/s sites that had nearly identical chemical compositions. Their interpretation is that different physical conditions during the formation of the illite were responsible for the structural variations.

The observation that Al-rich I/S from K-bentonites is dominated by c i s -vacan t layers does not have a clear crystallographic explanation although some specula- t ion may be worthwhile. Preliminary evidence indi- cates that the cv- 1 M structure is more stable than the t v - l M type (Drits et a l 1993). In the c v - l M structure, the six K-O bond lengths are equal, whereas in the tv- 1 M structure K-O bond lengths are unequal, three being shorter and three being longer. In the case of a t rans -occup ied octahedra, the geometric center of the distorted hexagonal ring created by the basal oxygens of the top tetrahedral sheet is shifted < a / 3 from the center of the corresponding ring on the bottom sheet (Tsipursky and Drits 1984, Drits et a l 1993). In t rans- vacant octahedra the geometric centers are shifted > 0,/3 from each other. In t r ans - vacan t muscovite-2M~ the interlayer K is not in the exact center of the distorted hexagonal ring but is displaced along the two-fold axis toward the vacant octahedral sites above and below resulting in unequal K-O bond lengths (Bailey 1984). Similar K-O bond length differences are likely for trans- vacant 1 M structures. There is less distortion of the tetrahedral sheets in the c i s -vacan t 1 M structure making a more regular interlayer cavity that results in nearly equal K-O bond lengths. Such differences in K bonding may affect the stability of these structures. How- ever, it is still not clear at this time why Mg and Fe should favor the tv structure. The cv structure may be the most stable form of US for high Al and relatively low Mg compositions. However, more work is needed to verify this hypothesis. Geological considerations and associations concerning cv and tv I/S from K-bentonites will be given in a future paper.

CONCLUSIONS

The three-dimensional crystal structure of disordered I/S can be quantified by modeling random-powder X-ray diffraction patterns using the computer algorithm WILDFIRE (Reynolds 1993). Previously, diffraction patterns from rotationally disordered I/S that were strikingly different were described simply as 1Md. The use of this modeling technique has enabled quan- tification of the amount and type of rotational disorder

and the distribution of octahedral cations in the I/S examined here.

In the US from Ordovician and Devonian K-bentonites of the eastern United States, cv interstratifica- t ion correlates with high tetrahedral A1, high octahedral A1 and n. 120 ~ rotational disorder. I/S that is dominated by tv octahedra correlate with low tetrahedral At substitution, high octahedral Mg and Fe substitution and n-60 ~ rotational disorder. I/S expandability and Reichweite does not correlate systematically with cv interstratification and rotational disorder. The structural features in I/S result from some combinat ion of their crystal chemistry and the environmental conditions under which the I/S formed.

ACKNOWLEDGMENTS

This research was supported in part by the American Chemical Society Grant 23613-AC2; Dartmouth Col- lege Earth Sciences Department, Hanover, NH; and the United States Geological Survey, Water Resources Division, Boulder, CO. Many of the samples in this study were generously donated by W. C. EIliott, W. D. Huff, R. L. Hay, and A. Chagnon who deserve special thanks. In addition, the writers are indebted to J. W. Delano, J. B. Droste, and S. P. Altaner for also sup- plying samples. The authors are indebted to Dr. Victor Drits for his review of this manuscript and elucidation concerning the structure of c/s-vacant and t rans -vacan t unit-cells. We would also like to thank Dr. Dennis Eberl for help and support and the referees and editors for valuable comments.

REFERENCES

Altaner, S. P., and C. M. Bethke. 1988. Interlayer order in illite/smectite. American Mineralogist 73: 766-774.

Bailey, S.W. 1966. The status of clay mineral structures. Clays & Clay Miner. 14: 1-23.

Bailey, S.W. 1980. Structures of layer silicates. In Crystal Structures o f Clay Minerals and their X-Ray Identification. G. W. Bilndley and G. Brown, eds. London: Mineralogical Society, 1-124.

Bailey, S.W. 1984. Crystal chemistry of the true micas, Chapter 2. In Micas. S. W. Bailey, ed. Reviews in Miner- alogy Vol. 13, Blacksburg, Virginia: Mineralogical Society of Ameilca, 13-60.

Bethke, C. M., and S. Marshak. 1990. Brine migrations across North American--The plate tectonics of ground- water. Annu. Rev. Earth Planet Sci. lg: 287-315.

Dennison, J. M., and D. A. Textoris. 1970. Devonian Tioga tuffin northeastern United States. Bulletin Volcanogenique 34: 289-293.

Dolasse, W.A. 1986. Correction of intensities for preferred orientation in powder diffractometry. Application of the March model. J. Appl. Cryst. 19: 267-272.

Dills, V. A., B. A. Plan~on, B. A. Sakharov, G. Besston, S. I. Tsipursky, and C. Tchoubar. 1984. Diffraction effects calculated for structural models of K-saturated montmo- rillonite containing different types of defects. Clay Miner. 19: 541-561.

Drits, V. A., and C. Tchoubar. 1990. X-ray Diffraction by Disordered Lametlar Structures. New York: Springer-Ver- lag, 371 pp.

Drits, V. A., F. Weber, A. L. Salyn, and S. I. Tsipursky. 1993.


X-ray identification of one-layer illite varieties. Application to the study of illites around uranium deposits of Canada. Clays & Clay Miner. 41: No. 3, 389-398.

Drits, V. A., and D. K. McCarty. 1995. The nature of diffraction effects from illite and illite/smectite consisting of interstratified trans-vacant and cis-vacant 2:1 layers; A semi- quantitative technique for determination of layer-type content. Amer. Miner. (in review).

Droste, J. B., and C. J. Vitiliano. 1973. Tioga bentonite (Middle Ordovic ian)of Indiana. Clays & Clay Miner. 21: 9-13.

Elliott, W. C., and J. L. Aronson. 1987. Alleghanian episode of K-bentonite iUitization in the southern Appalachian basin. Geology 15: 735-739.

Elliott, W. C., and J. L. Aronson. 1993. The timing and extent ofillite formation in Ordovician K-bentonites at the Cincinnati Arch, Nashville Dome and north-eastern Illinois basin. Basin Research 5: 125-135.

Giiven, N. 1971. Structural factors controlling stacking sequences in dioctahedral micas. Clays & Clay Miner. 134: 159-165.

Harris, A. G. 1979. Conodont color alteration, an organo- mineral metamorphic index and its application to Appa- lachian Basin geology. In Aspects ofDiagenesis. P. A. Scholle and P. R. Schluger, eds. Society of Economic Paleontologists and Mineralogists Special Publication 26:3-16.

Heam, P. P., J. F. SuRer, and H. E. Belkin. 1987. Evidence for Late-Paleozoic brine migration in Cambrian carbonate rocks of the central and southern Appalachians. Implica- tions for Mississippi Valley-type sulfide mineralization. Geochim. Cosmoch. Acta 51: 1323-1334.

Hower, J., E. V. Eslinger, M. E. Hower, and E. A. Perry. 1976. Mechanism of burial metamorphism of argillaceous sedi- ment, mineralogical and chemical evidence. Geol. Soc. of Amer. Bull. 87: 725-737.

Huff, W. D., and A. G. Tiirkmenoglu. 1981. Chemical char- acteristics and origin of Ordovician K-bentonites along the Cincinnati Arch. Clays & Clay Miner. 29:113-123.

Jennings, S., and G. R. Thompson. 1986. Digenesis of Plio- Pleistocene sediments of the Colorado River Delta, southern California. J. Sed. Petrology 56: 89-98.

Johnsson, M . J . 1984. The thermal and burial history of south central New York: Evidence from vitrinite reflec- tance, clay mineral diagenesis and fission track dating of apatite and zircon: Masters thesis. Dartmouth College, Hanover, New Hampshire, 155 pp.

Kolata, D. R., J. K. Frost, and W. D. Huff. 1984. K-bentonites of the Ordovician Decorah Subgroup, upper Mis- sissippi Valley: Correlation by chemical fingerprinting. II- finois State Geological Survey, Circular 537, 30 pp.

Maxwell, D. T., and J. Hower. 1967. High-grade diagenesis and low-grade metamorphism of illite in the Precambrian Belt Series. Amer. Miner. 52: 843-857.

McCarty, D. K., and G. R. Thompson. 1991. Burial diagenesis in two Montana Tertiary basins. Clays & Clay Min- er. 39: 293-305.

Mrring, J. and A. Oberlin. 1967. Electron-optical study of smectites: Clays & Clay Miner. 17th Nat. Conf., Pergamon Press, 3-25.

Mitchell, C. E. 1992. Chronostratigraphy of the Trenton Group and Utica Shale, Pt. I: Preliminary revision of lith- oracles and age relationships. Abstracts with Programs, 1992 GSA Annual Meeting, Cincinnati, Ohio.

Moore, D. M., and R. C. Reynolds Jr. 1989. X-ray Dif- fraction and the Identification and Analysis of Clay Min- erals. New York: Oxford University Press, 332 pp.

Oliver, J. 1986. Fluids expelled tectonically from orogenic belts: Their in hydrocarbon migration and other geologic phenomena. Geology 14: 99-102.

Perry, E., and J. Hower. 1970. Burial diagenesis in Gulf Coast pelitic sediments. Clays & Clay Miner. 18: 165-177.

Planqon, A. 1981. Diffraction by layer structures containing different kinds of layers and stacking faults. J. Appl. Cryst. 14: 300-304.

Plan~on, A., and C. Tchoubar. 1977a. Determination of structural defects in phyllosilicates by X-ray powder dif- f r ac t ion- I. Principle of calculation of the diffraction phe- nomenon. Clays & Clay Miner. 25: 430--435.

Plan~on, A., and C. Tchoubar. 1977b. Determination of structural defects in phyUosilicates by X-ray powder diffraction--II. Nature and proportion of defects in natural kaolinite. Clays & Clay Miner. 25: 436--450.

Plan~on, A., R. F. Giese, and R. Snyder. 1988. The Hinckley index for kaolinites. Clay Miner. 23: 249-260.

Radoslovich, E. W. 1959. Structural control of polymor- phism in micas. Nature 183: 253-254.

Radoslovich, E. W., and K. Norrish. 1962. The cell dimensions and symmetry of layer lattice silicates. I. Some structural considerations. Arner. Miner. 47:599-616.

Reynolds, R. C. 1985. NEWMOD computer program for the calculation of the one-dimensional X- ray diffraction patterns of mixed-layer clays. R. C. Reynolds, ed. Dept. of Earth Sciences, Dartmouth College, Hanover, New Hamp- shire 03755.

Reynolds, R . C . 1992. X-ray diffraction studies of illite/ smectite from rocks, < 1 ~tm randomly oriented powders, and < 1 ~tm oriented powder aggregates: The absence of laboratory-induced artifacts. Clays & Clay Miner. 40: 387- 396.

Reynolds, R. C. 1993. Three-dimensional powder X-ray diffraction from disordered illite: Simulation and interpretation of the diffraction patterns. In CMS Workshop Lec- tures, Vol. 5. Computer Applications to X-ray Powder Dif- fraction Analysis of Clay Minerals. R. C. Reynolds Jr. and J. R. Walker, eds. Boulder, Colorado: The Clay Minerals Society, 43-77.

Reynolds, R. C., and C. H. Thomson. 1993. Iilite from the Potsdam Sandstone of New York: A probable noncentrosymmetric mica structure. Clays & Clay Miner. 41: 66-72.

Roden, M. K., D. S. Miller, W. C. Eiliott, and J. L. Aronson. 1992. The thermal history of the distal margin and interior of the southern Appalachian basin from combined fission- track and K/Ar studies of K-bentonites (abstract). 27th An- nual Northeastern Section of the Geological Society of A mer- ican Meeting 24, no. 3, p. 49.

Roden, M. K., W. C. Elliott, J. L. Aronson, and D. S. Miller. 1993. A comparison of fission-track ages of apatite and zircon to the K/Ar ages of illite/smectite (I/S) from Ordo- vician K-bentonites, southern Appalachian basin. Journal of Geology 101:633-641.

Sakharov, B. A., G. Besson, V. A. Drits, M. Yu Kamenava, A. L. Salyn, and B. B. Smoliar. 1990. X-ray study of the nature of stacking faults in the structure ofglauconites. Clay Miner. 25: 419--435.

Srodofi, J., D. J. Morgan, E. V. Eslinger, D. D. Eberl, and M. R. Kadinger. 1986. Chemistry ofillite/smectite and end- member illite. Clays & Clay Miner. 34: 368-378.

Tsipursky, S. I., and V. A. Drits. 1984. The distribution of octahedral cations in the 2:1 layers ofdioctahedral smectites studied by oblique-texture electron diffraction. Clay Miner. 19: 177-193.

Velde, B. 1965. Experimental determination of muscovite polymorph stabilities. Amer. Miner. 50: 436-449.

Yoder, H. S., and H. P. Eugster. 1955. Synthetic and natural muscovites. Geochim. Cosmochim. Acta 8: 225-280.

(Received 8 February 1994; accepted 29 September 1994; Ms. 2467)

Documents

ROTATIONALLY DISORDERED ILLITE/SMECTITE IN PALEOZOIC …clays.org/journal/archive/volume 43/43-3-271.pdf · Vol. 43, No. 3, 1995 RotationaUy disordered illite/smectite in Paleozoic